JP5153660B2 - Image processing apparatus, image processing method, and program - Google Patents

Description

  The present invention relates to an image processing technique for detecting a human skin color region from an image.

  With the development of multimedia in recent years, various video compression encoding systems have been proposed, and representative ones include MPEG-1, 2, 4 and H.264. There are 26L. When image compression / encoding is performed in conformity with the MPEG system, the amount of code often varies greatly depending on the spatial frequency characteristics, scenes, and quantization parameters that are characteristics of the image itself. An important technique for obtaining a decoded image with good image quality in realizing such coding characteristics is code amount control and quantization control, and a method called TM5 (Test Model 5) is used. Commonly used.

  The TM5 method is roughly divided into three steps. In step 1, a target code amount of a picture to be encoded is determined. Step 2 sets the reference value of the quantization parameter for the macroblock so that the actual generated code amount approaches the target code amount in the picture.

  Step 3 executes a process of determining a quantization parameter based on the spatial activity of the macroblock to be encoded so that the visual characteristics, that is, the quality of the decoded image is good. Specifically, the spatial activity in the 8 × 8 sub-block in the macroblock is calculated, and the quantization parameter used for quantization is determined by multiplying the reference value of the quantization parameter calculated in step 2. By this processing, a large amount of code is allocated to a flat portion (low spatial activity) where visual degradation is conspicuous, so that it is possible to realize code amount control and quantization control with suppressed image quality degradation.

  On the other hand, a method for improving the image quality of the gaze area by incorporating a process of detecting the skin color of the person and assigning a large amount of code to the detected area as the skin color is proposed in Step 3.

  Many of the skin color detection methods are detection methods using threshold values, and various proposals have been made in the past. The reason is that it is known that the skin color of a person exists in a specific area in the color area. An example of skin color detection will be described below.

  As a general skin color detection method, there is a method of calculating an average value of color differences of pixels included in a block of a predetermined size and determining that the block is a skin color when included in a threshold (th0 to th1). .

  As another skin color detection method, it is determined whether the color difference value of each pixel in the block is within a predetermined threshold (th2 to th3), and pixels in th1 to th2 are set to another threshold (th4). There is a method for determining that the block contains a lot of skin color when the value exceeds. This block may be a macroblock using consistency with encoding (quantization).

As another method, color information is converted from RGB space to HI (V) space, and skin color is extracted with a threshold value determined in advance by H and I (see Patent Document 1). There are white balance setting and skin care treatment by the camera.
JP 2000-048184 A

  However, in skin color detection using the above threshold, it is difficult to accurately detect the skin color of a person due to the influence of lighting conditions, climate change, and the like. Since the optical signal input to the imaging device is weak in a dark room or on a cloudy day, signal amplification (gain increase) is performed inside the device, so that there may be a color space different from the actual skin color.

  In the natural world, there are many objects having a color space that is close to the skin color of a person, even though it does not look like a skin color to the eyes of a person, such as a wood desk or autumn leaves. For this reason, an area that is originally intended to be detected as skin color cannot be detected, or a non-skin color area is erroneously detected as a skin color, so that a larger amount of code is assigned to a non-skin color area than a skin color area, resulting in a reduction in image quality. There is also a risk of causing.

  The present invention has been made in view of the above problems, and improves the accuracy of skin color detection and achieves high image quality.

In order to solve the above-described problem, an image processing apparatus according to the present invention detects a specific color area included in an image, and the code amount of an area where the code amount of the specific color area is not the specific color area. an image processing apparatus that turn into symbols as obtain increased than an image input means for inputting images, face detection for detecting an area corresponding to a person's face from the input image by the image ZoIri force means means and a feature extraction means for extracting color information of the region corresponding to the face of the detected person, the specific color area detecting means for detecting a region of a specific color from the image image, color information of the extracted based on, it has a setting means for setting a threshold value for detecting the region of the specific color by the specific color area detection unit, corresponding to the face of a plurality of persons from the image by the face detecting unit more If this area is detected,
The setting means sets a threshold for detecting the specific color area by the specific color area detection means based on color information extracted from the area corresponding to each detected human face, The specific color area detecting means detects the specific color area for each area corresponding to the detected human face based on the threshold .

The image processing method according to the present invention detects the region of a specific color included in the image, such that the code amount of the specific color region is obtaining increased than the code amount of the non-specific color area region an image processing method that turn into symbols, a face detection step of detecting an image input step of inputting images, an area corresponding from the front Kiga image on the face of a person, corresponding to the face of the detected person a feature amount extraction step of extracting color information of the region, and the specific color area detection step of detecting a region of a specific color from the image image, based on the extracted color information, specified by the specific color area detection step possess a setting step of setting a threshold value for detecting the area of color, and when a plurality of regions corresponding to a person's face from the image in the face detection step is detected in the setting step, the detection The area corresponding to the face of each person Based on the extracted color information, a threshold value for detecting the specific color region is set by the specific color region detection step. In the specific color region detection step, each of the detected values based on the threshold value is set. The specific color area is detected for each area corresponding to a human face .

  According to the present invention, the skin color detection accuracy can be improved and the image quality can be improved by dynamically changing the threshold value used for the skin color detection using the face feature amount obtained from the face detection result.

    The best mode for carrying out the present invention will be described below in detail with reference to the accompanying drawings.

  The embodiment described below is an example for realizing the present invention, and should be appropriately modified or changed according to the configuration and various conditions of the apparatus to which the present invention is applied. It is not limited to the embodiment.

[First Embodiment]
First, a first embodiment according to the present invention will be described with reference to FIGS.

  FIG. 1 is a block diagram showing the overall configuration of the image processing apparatus according to the first embodiment of the present invention. FIG. 2 is a flowchart showing an image processing method by the image processing apparatus of the present embodiment. FIG. 3 is a diagram for explaining the face detection process and the skin color detection process. FIG. 4 is a diagram illustrating an image whose face is not detected. FIG. 5 is a diagram for explaining a skin color detection area setting method.

  The image processing apparatus of the present embodiment uses MPEG (Moving Pictures of Experts Group) or H.264 as an encoding method. This is a digital video camera to which H.264 / AVC (Advanced Video Coding) is applied. In addition, the image processing apparatus according to the present embodiment has a function of detecting a flesh color area included in an image, increasing the code amount of the flesh color area, and performing a coding process with high image quality.

  In FIG. 1, reference numerals 100 to 110 are blocks that mainly constitute a camera processing system, and reference numerals 120 to 137 are blocks that mainly constitute a recording processing system.

  Among the camera processing systems, an imaging optical system 100 including a focus lens forms an optical image of a subject on a light receiving surface of an image sensor 101 such as a CCD or CMOS, and photoelectric conversion is performed by the image sensor 101 to generate an image signal. Is done. The image signal output from the image sensor 101 is subjected to necessary signal processing such as color separation, gamma correction, white balance adjustment and the like by the camera signal processing unit 102 to form a luminance / color difference signal and store it in the frame memory 103. . The luminance / color difference signal stored in the frame memory 103 is read out by an image scaling unit 104, and an image range is read out by a predetermined pixel interpolation process, and a predetermined recording format is used. Is converted to

  The camera processing system is provided with a face detection unit 106, and either the output signal from the camera signal processing unit 102 or the signal converted into a predetermined recording format by the image scaling unit 104 is input by the first selector 105. Input selectively.

  The first selector 105 is for switching input depending on the performance of the face detection unit 106, and it is assumed that the image size converted by the image scaling unit 104 is smaller than the output image size of the camera signal processing unit 102. . When the processing performance of the face detection unit 106 is sufficiently high or when performing face detection with high accuracy, the output signal from the camera signal processing unit 102 is input to the first selector 105 in order to directly use the input image for face detection. To do. On the contrary, when the processing performance of the face detection unit 106 is not sufficient or when high accuracy is not required, the signal converted into a predetermined recording format by the image scaling unit 104 is input to the first selector 105. In this embodiment, either signal may be selected, and the number of faces to be detected may be dynamically switched, and the present invention is not particularly limited with respect to this selection method.

  When a face is detected by the face detection unit 106, the detected face size and position information is output to the face feature amount extraction unit 109. The face feature quantity extraction unit 109 extracts the detected face color information and outputs it to the camera system controller 110. The face size and position information is also output to the camera controller 110 and used for setting the skin color detection area. Details of the processing here will be described later together with the skin color detection unit 135 of the recording processing system in the subsequent stage.

  The detection result of the face detection unit 106 is output to the lens control unit 107 of the camera processing system, and the imaging optical system 100 is controlled to focus on the face area. Similarly, the face detection result is output to the camera control unit 108, and exposure correction and skin beautification processing for improving the image quality of the face are also executed. However, this function is not essential for realizing the present invention.

  Next, of the recording processing system, a signal converted into a predetermined recording format by the camera processing system is held in the frame memory 120 which is an input unit of the recording processing system. The signals held in the frame memory 120 are rearranged according to the encoded picture type by the frame rearrangement unit 121 in an order compliant with a predetermined encoding format. The image that has been rearranged by the frame rearrangement unit 121 is divided into blocks. Thereafter, when the encoded picture is an intra-frame coding (intra coding) system, the orthogonal transform unit 123 performs orthogonal transform processing on the signal in the macroblock, and the quantization unit 124 determines the orthogonal transform coefficient. Quantized. Further, when the encoded picture is an inter-frame encoded picture (inter-encoded) system, the inverse quantization unit 125 performs inverse quantization on the already encoded picture, and the inverse orthogonal transform unit 126 performs inverse orthogonality. A locally decoded image that has been subjected to the conversion process is created.

  Next, the motion prediction / motion compensation unit 127 performs motion prediction and motion compensation with the picture to be encoded, and the difference value with respect to the local decoded image is calculated by the addition / subtraction unit 122. The calculated difference value is subjected to orthogonal transformation processing by the orthogonal transformation unit 123, and the orthogonal transformation coefficient is quantized by the quantization unit 124.

  Regardless of intraframe coding or interframe coding, the quantized signal quantized by the quantizing unit 124 is encoded by the variable length encoding unit 131, and the encoded signal is temporarily held by the stream buffer 132. Thereafter, the data is recorded on a recording medium 136 such as an HDD or an optical disk.

  On the other hand, the code amount control unit 133 allocates an allocation bit amount for each picture in the GOP according to a bit amount for a picture that has not been encoded in the GOP including the allocation target picture. This allocation is repeated in the order of the encoded pictures in the GOP, and a picture target code amount is set for each picture.

  The quantization control unit 134 calculates a quantization parameter by feedback control based on the capacity of the virtual buffer for matching the target code amount for each picture with the generated code amount output from the variable length encoding unit 131. Then, the macro block detected as the skin color by the skin color detection unit 135 is controlled to be smaller than the calculated quantization parameter. By reducing the quantization parameter, the amount of code increases in the macroblock detected as flesh color, resulting in high image quality.

  Next, the face detection process and the skin color detection process in this embodiment will be described with reference to the flow of FIG. 2 and FIGS.

  In FIG. 2, when power is supplied from an AC power source or a battery and the image processing apparatus is activated, first, in S201, the value of M, which is the number of frames in which the face detection unit 106 does not continuously detect a face, is set to zero. Reset to. At this time, each block of the apparatus including the camera processing system and the recording processing system shifts to the recording standby state.

  In S <b> 202, as an image input process, an image is input to the face detection unit 106 via the imaging optical system 100, the image sensor 101, and the camera signal processing unit 102. The flow of image signals other than the face detection unit 106 is as described above.

  In S203, the face detection unit 106 performs face detection processing. If a face is detected in S203, the process proceeds to S204. If a face is not detected, the process proceeds to S212, but the process here will be described later.

  In S204, since the face is detected in the current frame, the number M of frames in which no face is continuously detected is reset to zero.

  In S205, a skin color detection area (AM0) for performing skin color detection from the face size and position information is set for the area detected as a face by the face detection unit 106. This setting method will be described with reference to FIG.

  In FIG. 5, reference numeral 501 denotes an input image, and reference numeral 502 denotes an area detected as a face by the face detection unit 106. Reference numeral 503 denotes an image in which a skin color detection area is set based on face size and position information with respect to the input image 501. Reference numeral 504 denotes an area setting process, and reference numeral 505 denotes a skin color detection area.

  From the size and position information of the area 502 detected as a face, a skin color detection area is set downward with respect to the input image 501 with reference to the face position. In general, since the human torso is wider in the horizontal direction than the face, an area below the face is set wider in the horizontal direction than the skin color detection area 504 as in 505. Note that the shape of the skin color detection region is not limited to a quadrangle such as 505 but may be a trapezoid such as 506, and can be arbitrarily set according to the size of the face. Further, if it has a function of detecting a human torso from the face detection result, a skin color detection region may be set for the torso. Further, the present invention is not limited to the above-described skin color detection area setting method.

  Returning to the description of FIG. 2, in step S <b> 206 a, the face feature amount extraction unit 109 extracts color information included in the face from the area detected as a face by the face detection unit 106. In step S <b> 206 b, the face feature amount extraction unit 109. The threshold value used for the skin color detection unit 135 is set from the extraction result of. The color information here is color space information such as YUV (YCbCr), RGB, and HSV.

  Here, face feature amount extraction processing and skin color detection processing using YCbCr as an example of the color space will be described with reference to FIG.

  In FIG. 3, reference numeral 301 denotes an input image. As a result of performing face detection on the input image 301, it is assumed that a face is detected in the face detection area 302 as in 501 and 502 of FIG. 5. Next, the Cb value (peak value) occupying the largest number of pixels in the Cb value histogram included in the face detection area 302 is defined as Cbtop. Cbtop is considered to be the color closest to the skin color constituting the face in the face detection area 302. From this, it is considered that not only the face included in the face detection area 302 but also the skin color of the person's neck, hands, feet, etc. is centered on Cbtop. Therefore, CbLM0 and CbHM0 are set as threshold values of the Cb value used in the skin color detection unit 135, assuming that the Cb value within a predetermined range from Cbtop is the skin color of this person (303). The predetermined range here can be arbitrarily determined. Also, CrLM0 and CrHM0 are set as threshold values for the Cr value as well as the Cb value (303). In addition, although YCbCr was illustrated as a color space, in the case of another color space, a threshold value corresponding to the space may be set.

  Next, the process proceeds to S207, and skin color detection is performed on the skin color detection region set from the face size and position information using the skin color detection threshold set in S206b (304 in FIG. 3). In addition, skin color detection is not performed for regions other than the skin color detection region. This makes it difficult for erroneous detection of the skin color in an area where there is a high possibility that the skin color does not exist.

  On the other hand, if a face is not detected in S203, the process proceeds to S212, and M is updated by adding 1 to M, which is the number of frames for which no face is detected.

  In S213, if the number M of frames for which no face is detected is shorter than the predetermined period T0, the process proceeds to S214a, and if it is longer than T0, the process proceeds to S215.

  In S214a and 214b, the skin color detection region and the skin color detection threshold value at the time of the most recent face detection (that is, M = 0) are set, and the process proceeds to S207 to perform skin color detection. This is because when a person temporarily turns sideways or down like 401 in FIG. 4 or when the face is not detected for a moment due to the processing amount of the face detection unit 106, etc. This is because the person itself is considered to exist in the image.

  On the other hand, when the number M of frames in which no face is detected is longer than the predetermined period TO, the process proceeds to S215 and skin color detection is not performed. This is because a person's face is not detected for a moment as in 402 of FIG. 4 and it is considered that no person exists in the image.

  In S208, when the face is detected or the number M of frames in which no face is detected is shorter than the period TO, the skin color is detected according to the skin color detection threshold set in S206b or S214b, and the input is made based on the result. Encode the image. Here, the quantization control unit 134 determines the quantization parameter so that the code amount of the skin color region is allocated more than the region other than the skin color region.

  Next, in S209, the encoded image signal is stored in the recording medium 136. In S210, when the recording is not finished, that is, when the recording is continued as it is, the process proceeds to S211. In S211, the current skin color detection threshold value is retained, and then the process proceeds to S202, and the processing of S202 and subsequent steps is repeatedly executed in response to the input of the next frame image.

  On the other hand, if the recording is finished in S210, the process is finished.

[Second Embodiment]
In the second embodiment, there are a plurality of persons in the input image.

  The configuration of the image processing apparatus according to the present embodiment is the same as that shown in FIG.

  FIG. 6 is a diagram for explaining an image processing method according to the second embodiment, in which a plurality of persons are included in the image of FIG.

  In FIG. 6, reference numeral 601 denotes an input image, and it is assumed that a person A and a person B exist. Reference numeral 602 denotes an area where the face detection unit 106 detects a face of the person A, and reference numeral 603 denotes a skin color detection area of the person A set from the face size and position information with respect to the input image 601. Similarly, reference numeral 604 denotes an area in which the face detection unit 106 detects a face for the person B, and reference numeral 605 denotes a skin color detection area for the person B set from the face size and position information for the input image 601. .

  By the same method as in the first embodiment, a skin color detection area for the person A is set for the input image 601 from the size and position information of the area 602 detected for the face of the person A as in 603. Similarly, a skin color detection area for the person B is set as indicated by 605 from the size and position information of the face of the person B and the detected area 604.

  Next, the face feature amount extraction unit 109 extracts color information included in the face of the person A, and sets a skin color detection threshold for the skin color detection region 603. Similarly, for the face of the person B, a threshold value used for skin color detection is set in the skin color detection area 605. Each setting method of the skin color detection area and the threshold is the same as that in the first embodiment, but is not limited thereto.

  Next, skin color detection is performed on the skin color detection region set from the face size and position information using the skin color detection threshold set from the face skin color information (606). In addition, skin color detection is not performed for areas other than the skin color detection area. Thereby, it becomes difficult to generate the false detection of the skin color in an area where there is a high possibility that the skin color does not exist.

  Other processes are the same as those in the first embodiment. Further, in the second embodiment, two persons are exemplified, but three or more persons may be used and depend on the number of faces that can be detected by the face detection unit 106.

[Third Embodiment]
The third embodiment is a case where a plurality of persons exist in the input image and the skin color detection areas set from the face detection results overlap each other.

  The configuration of the image processing apparatus according to the present embodiment is the same as that shown in FIG.

  FIG. 7 is a diagram for explaining an image processing method according to the third embodiment, in which skin color detection regions overlap for a person in the image of FIG.

  In FIG. 7, reference numeral 701 denotes an input image, and it is assumed that a person A and a person B exist. Reference numeral 702 denotes an area where the face detection unit 106 detects a face for the person A, and reference numeral 703 denotes a skin color detection area for the person A set from the face size and position information for the input image 701. Similarly, 704 is an area where the face detection unit 106 detects a face for the person B, and 705 is a skin color detection area for the person B set from the face size and position information for the input image 701. . Reference numeral 706 denotes an area where the skin color detection area 703 of the person A overlaps with the skin color detection area 705 of the person B.

  Similar to the first embodiment, a skin color detection area for the person A is set for the input image 701 from the size and position information of the area 702 detected for the face of the person A as in 703. Similarly, a skin color detection area for person B is set as indicated by 705 from the size and position information of the face of person B and the detected area 704.

  Next, the face feature quantity extraction unit 109 extracts color information included in the face of the person A, and sets a skin color detection threshold for the skin color detection area 703. Similarly, for the face of the person B, a skin color detection threshold used for skin color detection is set for the skin color detection region 705. The skin color detection area and the threshold value setting method are the same as those in the first embodiment, but are not limited thereto.

  Next, a method for setting the skin color detection threshold for the overlapping region 706 will be described.

  The first method will be described. Reference numeral 711 denotes a skin color detection threshold value set in the skin color detection area 703 of the person A, 712 denotes a skin color detection threshold value set in the skin color detection area 705 of the person B, and 713 denotes a skin color detection threshold value set in the overlap area 706. Yes. The Cb value is used as the color information. In this case, since the threshold 711 set in the skin color detection area 703 is highly likely to detect a skin color in a wider range than the threshold 712 set in the skin color detection area 705, the skin color detection threshold set in the overlap area 706 is , 712 having a detection range wider than 712.

  The second method will be described. When comparing the skin color detection area 703 of the person A and the skin color detection area 705 of the person B, the skin color detection area 703 of the person A is wider. In this case, since the overlapping area 706 is likely to be occupied by the skin color of the person A, the threshold value set in the overlapping area 706 is set to the same value as 711.

  A third method will be described. Reference numeral 721 denotes a skin color detection threshold set in the skin color detection area 703 of the person A, 722 denotes a skin color detection threshold set in the skin color detection area 705 of the person B, and 723 denotes a skin color detection threshold set in the overlap area 706. Yes. The Cb value is used as the color information. In this case, the skin color detection threshold 723 for the overlap region 706 is set so as to include both the threshold 721 set in the skin color detection region 703 and the threshold 722 set in the skin color detection region 705. This increases the possibility that the skin color of both the person A and the person B can be detected.

  The method for setting the skin color detection threshold for the overlap region and the method for setting the skin color detection threshold for each skin color detection region are not particularly limited to the methods described above.

  Further, in the third embodiment, two persons are exemplified, but three or more persons may be used, and depend on the number of faces that can be detected by the face detection unit 106.

[Other Embodiments]
In addition, each process in each embodiment mentioned above may implement | achieve the function by reading the program for implement | achieving the function of each process from memory, and CPU (central processing unit) of a computer executing. Good.

  The memory accessed by the CPU includes a non-volatile memory such as an HDD, an optical disk, and a flash memory, a recording medium such as a CD-ROM that can only be read, and a volatile memory other than the RAM. Or you may comprise from the computer-readable recording medium by these combination, and a writable recording medium.

  In addition, a program for realizing the function of each process in each of the above-described embodiments is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into a computer system and executed. Processing may be performed. Here, the “computer system” includes an OS and hardware such as peripheral devices. Specifically, the program read from the storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. After that, the CPU of the function expansion board or function expansion unit performs part or all of the actual processing based on the instructions of the program, and the functions of the above-described embodiments are realized by the processing.

The “computer-readable recording medium” refers to a portable medium such as an optical disk such as a CD-ROM or DVD, a semiconductor memory card, or a storage device such as a hard disk built in a computer system.
Further, the “computer-readable recording medium” includes a volatile memory (RAM) inside a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In other words, it includes those that hold programs for a certain period of time.

  The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.

  The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

  A program product such as a computer-readable recording medium in which the above program is recorded can also be applied as an embodiment of the present invention. The above program, recording medium, transmission medium, and program product are included in the scope of the present invention.

  As described above, the specific configuration is not limited to this embodiment, and includes design and the like within a range not departing from the gist of the present invention.

1 is a block diagram illustrating an overall configuration of an image processing apparatus according to a first embodiment of the present invention. It is a flowchart which shows the image processing method by the image processing apparatus of this embodiment. It is a figure explaining a face detection process and a skin color detection process. It is a figure which illustrates the image in which a face is not detected. It is a figure explaining the setting method of a skin color detection area. It is a figure explaining the image processing method of 2nd Embodiment. It is a figure explaining the image processing method of 3rd Embodiment.

DESCRIPTION OF SYMBOLS 100 Imaging optical system 101 Image pick-up element 102 Camera signal processing part 103 Frame memory (camera processing system)
104 Image scaling unit 105 First selector 106 Face detection unit 107 Lens control unit 108 Camera control unit 109 Face feature amount extraction unit 110 Camera processing system controller 120 Frame memory (recording processing system)
121 Frame Rearrangement Unit 122 Addition / Subtraction Unit 123 Orthogonal Transform Unit 124 Quantization Unit 125 Inverse Quantization Unit 126 Inverse Orthogonal Transformation Unit 127 Motion Prediction / Motion Compensation Unit 128 Addition / Subtraction Unit 129 Frame Memory (Recording Processing System)
130 Second selector 131 Variable length encoding unit 132 Buffer memory 133 Code amount control unit 134 Quantization control unit 135 Skin color detection unit 136 Recording medium 137 Recording processing system controller

Claims (26)

Detecting a region of a specific color included in the image, an image processing apparatus that turn into symbols as obtain increased than the code amount of area code amount of the specific color region is the non-specific color region,
An image input means for inputting images,
A face detecting unit for detecting an area corresponding to the face of a person from the input image by the image ZoIri force means,
Feature quantity extraction means for extracting color information of a region corresponding to the detected human face;
A specific color area detecting means for detecting a region of a specific color from the image image,
Based on the extracted color information, it has a setting means for setting a threshold value for detecting the region of the specific color by the specific color region detecting means,
When a plurality of regions corresponding to a plurality of human faces are detected from the image by the face detection means,
The setting means sets a threshold for detecting the specific color area by the specific color area detection means based on color information extracted from the area corresponding to each detected human face,
The image processing apparatus according to claim 1, wherein the specific color area detection unit detects the specific color area for each area corresponding to the detected human face based on the threshold . The face detecting unit detects the size and position of the region corresponding to the face of a person from the input image by the image input means,
The specific color region detecting means, wherein, characterized in that to determine the region for detecting the region of the specific color by using the size and position of the area your capital corresponding to the face of the detected person Item 8. The image processing apparatus according to Item 1. The specific color area detecting means is set when the area corresponding to the person's face is detected most recently when the period corresponding to the person's face is not detected by the face detecting means is shorter than a predetermined period. the image processing apparatus according to claim 1 or 2, characterized in that the detection of the specific color by using the threshold value. The color information, the image processing device according to claim 1, wherein in any one of the three that the color space information with pixels. The image processing apparatus according to claim 4 , wherein the threshold used for detecting the specific color is calculated from a peak value in a histogram of color space information included in an area corresponding to the face of the person . If the area for detecting an area of the specific color are overlapped, selects one of the threshold used to detect a specific color area of each detection area overlapping, the overlapping with the selected threshold value the image processing apparatus according to any one of claims 1 5, characterized in that the detection of a specific color region of the detection region. When said area for detecting a specific color area of the overlap, the detection range from the threshold used to detect a specific color region of the detection region overlapping selects the most widely made threshold, using the selected threshold value the image processing apparatus according to any one of claims 1-5, characterized in that the detection of a specific color region of the detection region of the overlap. If the area for detecting an area of the specific color are overlapped, the detection region from the threshold used to detect a specific color region of the detection region overlapping selects the threshold of the widest area, using the selected threshold value the image processing apparatus according to any one of claims 1 5, characterized in that the detection of a specific color region of the detection region of the overlapping Te. If the area for detecting an area of the specific color are overlapped, and sets a new threshold from the threshold used to detect a specific color area of each detection area overlapping, the overlapping with the threshold set above the newly the image processing apparatus according to claim 1, any one of 5, characterized in that the detection of a specific color region of the detection region to be. If the area for detecting an area of the specific color is overlapped, the detection range from the threshold used to detect a specific color region of the detection region overlapping sets the most widely made new threshold was set the new the image processing apparatus according to any one of claims 1 5, characterized in that the detection of a specific color region of the detection region of the overlap with a threshold. Imaging means for generating the images from the optical image of an object,
The recording apparatus according to claim 1, further comprising: a recording unit configured to perform encoding and recording so that a code amount for a specific color region in the captured image is larger than a code amount for a region that is not the specific color region. The image processing apparatus according to any one of 1 to 10 . Detecting a region of a specific color included in the image, the a specific color code amount the specific color area code turn into that image processing method as obtain increased than the code amount of area not in the areas of,
An image input step of inputting images,
A face detecting step of detecting an area corresponding from the front Kiga image to a person's face,
A feature amount extracting step of extracting color information of an area corresponding to the detected human face;
A specific color area detection step of detecting a region of a specific color from the image image,
Based on the extracted color information, have a, a setting step of setting a threshold for detecting a region of a specific color in the specific color area detection step,
When a plurality of regions corresponding to a human face are detected from the image in the face detection step,
In the setting step, based on the color information extracted from the region corresponding to each detected human face, a threshold for detecting the specific color region by the specific color region detection step is set.
In the specific color region detection step, the specific color region is detected for each region corresponding to each detected human face based on the threshold value . A program for causing a computer to execute the image processing method according to claim 12 .   Image acquisition means for acquiring images;
  Encoding means for encoding the image obtained by the image acquisition means,
  The encoding means determines a specific color area in the image obtained by the image acquisition means, and encodes the code amount of the specific color area to be larger than an area that is not the specific color area,
  The encoding means sets a threshold value for determining that the area is the specific color based on the color information of the area corresponding to the face of the person in the image obtained by the image acquisition means, and Determining the region of the specific color in the region corresponding to the face;
  When there are a plurality of regions corresponding to the person's face, the encoding means determines the first color region based on the color information of the region corresponding to the first person's face. And a second threshold value for determining that the region is the specific color region based on the color information of the region corresponding to the face of the second person. In the area for determining the corresponding specific color area, the specific color area is determined using the first threshold, and the specific color area corresponding to the face of the second person is determined. An image processing apparatus that determines the specific color area using the second threshold value in the area to be used.   The encoding means is configured to quantize a quantization parameter for the specific color area so that a code amount of the specific color area is larger than that for the non-specific color area. The image processing apparatus according to claim 14, wherein the image processing apparatus is smaller than the parameter.   The encoding means includes a region for determining a specific color region corresponding to the face of the first person and a region for determining a specific color region corresponding to the face of the second person. The image according to claim 14 or 15, wherein, in the overlapping region, the region of the specific color is determined using either one of the first threshold value or the second threshold value. Processing equipment.   The encoding means includes a region for determining a specific color region corresponding to the face of the first person and a region for determining a specific color region corresponding to the face of the second person. In this case, in the overlapping region, the region of the specific color is determined using a third threshold value including both the first threshold value and the second threshold value. The image processing apparatus described.   The encoding means includes a region for determining a specific color region corresponding to the face of the first person and a region for determining a specific color region corresponding to the face of the second person. 16. The method according to claim 14, wherein in the overlapping region, the region of the specific color is determined using a third threshold different from the first threshold and the second threshold. Image processing device.   The image processing apparatus according to claim 18, wherein the third threshold value includes both the first threshold value and the second threshold value.   The image processing apparatus according to claim 14, wherein the color information is color space information of a pixel.   21. The region for determining a specific color region corresponding to the person's face is determined based on a size and a position of the region corresponding to the person's face. The image processing apparatus according to item 1.   The first threshold value is calculated from a peak value in a histogram of color space information included in a region corresponding to the face of the first person. Image processing apparatus.   The image processing apparatus according to claim 14, wherein the image acquisition unit acquires an image by capturing an optical image of a subject.   24. The image processing apparatus according to claim 14, further comprising recording means for recording the image data encoded by the encoding means on a recording medium.   An image processing method for encoding an image obtained by an image acquisition means,
  Determining a specific color area in the image obtained by the image acquisition means, and encoding so that a code amount of the specific color area is larger than a non-specific color area;
  Based on the color information of the area corresponding to the face of the person in the image obtained by the image acquisition means, a threshold for determining that the area is the specific color is set, and in the area corresponding to the face of the person Determining the region of the specific color;
  When there are a plurality of regions corresponding to the person's face, a threshold value for determining that the region is the specific color based on the color information of the region corresponding to each person's face is set. An image processing method, wherein the specific color area is determined for each area corresponding to the face of the person.   A program for causing a computer to execute the image processing method according to claim 25.

Images